Re: AreaDetector OCR plugin

[Michael Davidsaver via Tech-talk <tech-talk@aps.anl.gov>]

On 2/15/19 7:20 AM, Rod Nussbaumer via Tech-talk wrote:
> All:
> 
> Has anyone developed an Optical Character Recognition plugin for AreaDetector? The potential use case is to have a camera aimed at an instrument front panel, and digitize the display to extract the instrument's displayed output. I've already anticipated the potential problems associated with multiplexed displays that may introduce artifacts into the digitized images, so I don't want to attempt to develop something that may be a failure anyway.

Do you have an OCR software in mind?

I investigated this a couple of years ago and the only functional FOSS
entrant seemed to be:

https://github.com/tesseract-ocr/tesseract

Which does an acceptable job with scans of printed documents
with small rotations.

> Are there any classes of inexpensive low resolution cameras supported by AreaDetector? Like consumer grade webcams, for instance?
> 
> Yes, it's a hack, but easier than the alternatives in at least one known case.

I look forward to hearing how this works out!